Predictors of 1-Year Global Outcomes After Traumatic Brain Injury Among Older Adults: A NIDILRR Traumatic Brain Injury Model Systems Study

Abstract

Objective: To assess predictors of global function and driving status among older adults (50 years and older) who survived 1 year following inpatient rehabilitation for moderate-to-severe traumatic brain injury (TBI). Methods: Functional status at 1-year post-TBI was determined for 1,845 individuals. The relationship age category to function was studied using associations and predictive modeling. Results: The final model accounted for 34% variance in Glasgow Outcome Scale-Extended (GOS-E) among 60- to 69-year-olds and 70- to 79-year-olds, and 25% variance in 50- to 59-year-olds and 80+-year-olds. FIM Motor at rehabilitation discharge made the greatest contribution to GOS-E variance across all age groups. Inpatient rehabilitation discharge to nursing home or adult home and rehospitalization were associated with a one-level decrease in GOS-E. Alcohol use predicted lower GOS-E among the 70- to 79-year-olds. Gender, ethnicity, and rehospitalizations were negatively associated driving. Discussion: Rehabilitation approaches to older adults with TBI may help maximize function and, thereby, improve later outcomes and decrease rehospitlaizations. Such strategies may include longer and more intensive acute rehabilitation with greater patient engagement and enhanced transitions of care.

Keywords

aging traumatic brain injury rehabilitation older adults

Introduction

By 2030, the United States population aged 65 and older is projected to be 72 million people which will represent nearly 20% of the total U.S. population (U.S. Department of Labor, Bureau of Labor Statistics, 2014; 2017 Profile of Older Americans). Rates of traumatic brain injury (TBI)-related hospitalizations and deaths are the highest among this cohort (Coronado, Thomas, Sattin, & Johnson, 2005). In 2012, annual hospital charges for treating older adult patients with TBI exceeded US$3.9 billion. Given how effective public health initiatives have been in improving life expectancy for older adults, it is prudent to investigate factors associated with improving outcomes in this cohort of older adult TBI survivors. This knowledge would better inform the medical and economic management of care for older adults with TBI as they enter the later decades of life.

Previous research regarding older adults with TBI has primarily focused on mortality and general functional outcomes (Coronado et al., 2005; Dams-O’Connor et al., 2013; Susman et al., 2002; Utomo, Gabbe, Simpson, & Cameron, 2009), often making broad comparisons between younger and older cohorts using variable age cut-offs to define older adulthood. These studies showed that not only do older adults tend to require greater frequency hospitalizations for milder TBI compared to younger adults (Dams-O’Connor et al., 2013), they also tend to have poorer outcomes following TBI even with similar lengths of hospital stay (Marquez de la Plata et al., 2008; Ponsford, Cameron, Fitzgerald, Grant, & Mikocka-Walus, 2011). Physiological and cognitive decline associated with aging present a unique challenge to understanding the long-term incremental effects of TBI in older adults. Health considerations such as preexisting chronic conditions, gradual cognitive impairment, and polypharmacy can complicate the course of recovery following injury.

Despite these age-specific considerations, there is generally a “one size fits all” approach to treat adults with TBI, especially beyond the acute care and acute rehabilitation phases (Gardner, Dams-O’Connor, Morrissey, & Manley, 2018; Thompson, McCormick, & Kagan, 2006). Care tailored more specifically to the needs of older adults with TBI may contribute to more strategic resource planning and health care delivery and may maximize quality of life and functional outcomes. A better understanding of factors associated with improved long-term outcomes after TBI, particularly factors that are modifiable, can inform the development of age-specific treatment guidelines for acute rehabilitation and long-term management among older adults.

Utilizing a large, longitudinal data set which has been weighted to represent the U.S. TBI inpatient rehabilitation population, the purpose of this study was to (a) characterize the demographic, clinical, and functional characteristics among older adults who receive inpatient rehabilitation for TBI and (b) evaluate the relationships between patient characteristics and level of functioning during inpatient rehabilitation (e.g., motor and cognitive burden of care, community participation, residence, alcohol use) with global outcomes and driving status at 1-year post-TBI across the older age groups. This study extends the findings of prior studies through the inclusion of those 50 and older, dividing the older adults into smaller age categories, and the use of outcome measures that are meaningful to those living with TBI.

Prior publications have utilized varied age cut-offs in the study of TBI among older adults, with the most common cut-points being 50, 60, and 65 years (Goleburn & Golden, 2001). People who are 50 years and older was chosen for the present study based on literature suggesting this age cohort has higher rates of comorbidities and complications, including deterioration in cognitive and emotional function (Goldstein, 1998; Goldstein & Levin, 1995; Goldstein et al., 1999; Goldstein et al., 1996; Kumar et al., 2019; Mazzucchi et al., 1992; Whiteneck et al., n.d.). In addition, including those between 50 and 65 years of age allows differences in payor sources.

About 1-year post-TBI, the first follow-up interval for the TBI Model Systems (TBIMS), was chosen as the outcome interval for this study for many reasons. One year is a well-established, well-studied postinjury interval, and therefore 1 year allows assessment consistent with prior literature. While 1 year is certainly not a time when people with TBI have plateaued in their recovery, the 1-year time point offers a glimpse of the landscape of outcomes after TBI. 1 year is also the earliest time point when most people post-TBI would be expected to have completed most of the treatment services (inpatient rehabilitation and outpatient therapy services). Thus, 1 year is a time point that may best reflect “rehabilitation outcomes.” Further out postinjury, outcomes would be expected to have greater potential to be influenced by the effects of aging, and also loss of participants to death. Finally, 1-year postinjury sets the stage for future studies that look longer postinjury.

Method

Participants

Data for this study were obtained from the TBIMS National Database (NDB). The TBIMS NDB, funded by the National Institute on Disability, Independent Living, and Rehabilitation Research, includes longitudinal data on individuals with TBI receiving inpatient rehabilitation at 20 TBIMS centers dating from 1987 to present day (TBIMS NDB, 2015). The TBIMS NDB defines TBI as

damage to brain tissue caused by an external mechanical force as evidenced by medically documented loss of consciousness or post-traumatic amnesia (PTA) due to brain trauma or by objective neurological findings that can be reasonably attributed to TBI on physical examination or mental status examination.

Participants must also meet the following inclusion criteria to be enrolled in the TBIMS NDB: (a) have a diagnosis of moderate to severe TBI as determined by one of the following—PTA longer than 24 hr, trauma-related intracranial neuroimaging abnormalities, loss of consciousness exceeding 30 min, or a Glasgow Coma Scale (GCS) score in the Emergency Department of less than 13 (unless due to intubation, sedation or intoxication); (b) be at least 16 years of age at the time of injury; (c) be admitted to a TBIMS’s acute care hospital within 72 hr of injury; (d) receive both acute hospital care and comprehensive rehabilitation in a designated brain injury program within the Model System; and (e) provide informed consent to participate or have a proxy consent. Each center received approval for human subjects research from their respective Institutional Review Boards (reference number 1211010085R005). Methods for enrollment, follow-up, data collection, and data management follow standard operating procedures. Data collectors undergo formalized training and interreliability testing. Demographic, premorbid, injury, and treatment-related data are abstracted from medical and administrative records, standardized patient assessments, and patient/family interview. Follow-up interviews are conducted by telephone or in-person using structured interviews where applicable (e.g., FIM, Disability Rating Scale, and Glasgow Outcome Scale-Extended). Test–Retest reliability of the TBIMS NDB follow-up interview revealed good to excellent test–retest reliability estimated with a few exceptions (exceptions were with variables not used in the present study; Bogner et al., 2017).

Data Source

Data collected by the Uniform Data System for Medical Rehabilitation (UDSMR; 2012) and the American Medical Rehabilitation Providers Association (eRehabData, 2012) provided national parameters with which the TBIMS NDB could be weighted to align with the national population (Corrigan et al., 2014; Cuthbert et al., 2014). Previous work by Corrigan et al. (2012) and Cuthbert et al. (2012) compared the TBIMS NDB to these data resources using a number of commonly collected variables, and demonstrated the representativeness of the TBIMS NDB in relation to the U.S. TBI inpatient rehabilitation population. Weighting the TBIMS NDB allows for the creation of national estimates of the numerous variables systematically collected as part of the TBIMS NDB during inpatient rehabilitation and long-term follow-up, and permits examination of the longitudinal effects of TBI in this population. The TBIMS NDB is particularly appropriate for this purpose as the database includes only medical record-verified cases of moderate-to-severe TBI that have been admitted to inpatient rehabilitation within a continuum of TBI care and has a high level of data accuracy due in part to extensive data quality procedures.

Weighting and Ranking Procedure

The TBIMS NDB was weighted to provide the U.S. TBI inpatient rehabilitation population estimates. The raking and weight trimming processes involved in weighting the data are described in detail elsewhere (Cuthbert et al., 2014). Briefly, the data were weighted to represent individuals ages 16 years and older with a primary diagnosis of TBI, who were admitted for inpatient rehabilitation and discharged between October 1, 2001, and December 31, 2010 yielding a weighted data set of 7,373 cases. Missing values for the weighted variables within the TBIMS NDB were rare, with less than 4% missing in any variable. For cases with any missing data, the values were imputed using expectation maximization (Dempster, Laird, & Rubin, 1977) prior to implementing the raking procedure. Following the raking procedure, weights were assessed using weight trimming, which truncates the weights of extreme values to less extreme values to reduce the mean squared error for any outcomes of interest. However, for these analyses, weight trimming was not found to reduce the mean squared error for any of the outcomes, so the weights were applied in their original configuration. All raking and weight trimming procedures were completed using SAS Version 9.3 (SAS Institute Inc., Cary, North Carolina). Differences in characteristics between individuals in the TBIMS NDB and the National Population being represented by the weighting have been described previously (Corrigan et al., 2012; Cuthbert et al., 2012).

Measures

Measures of interest were selected based on previously examined associations between covariates and the outcome measure (Brown et al., 2005; Graham et al., 2010) and also based on clinical experience. Variables selected for inclusion were categorized within four domains: sociodemographics, injury severity, functional status, and alcohol use.

Sociodemographics

Sociodemographic variables of interest included gender, race/ethnicity, education level, and primary rehabilitation payor source. Race/ethnicity was categorized as “Caucasian,” “African American,” “Hispanic/Latino origin,” and “Other,” which included Asian/Pacific Islander, Native American and other descent. Education was categorized into “11th grade or less,” “High School” (including HS diploma and GED), “Some College” (including associate’s degree and any work toward a bachelor’s degree), and “College and Postgraduate” (including bachelor’s, master’s, doctorate degrees and any graduate school work). Payor source for inpatient rehabilitation hospitalization comprised the following categories: private insurance (including private insurance, HMO, PPO, auto insurance, and TRICARE), Medicare, Medicaid, Worker’s compensation, and uninsured (no pay/hospital free care and self-pay).

Injury characterization

Variables used to characterize injury severity included cause of injury, GCS score at the time of Emergency Department admission, duration of Parent Teacher Association (PTA), and total length of stay (LOS) for acute care and inpatient rehabilitation. For these analyses, GCS was treated as a categorical variable with the following groupings: mild (13 to 15), moderate (9 to 12), severe (3 to 8), and sedated (no valid score due to clinical sedation). Total LOS, which is days from acute hospital admission to inpatient rehabilitation discharge, was also treated as a categorical variable with the following groupings: 1 to 9 days, 10 to 19 days, 20 to 29 days, 30 to 39 days, and 40+ days. This was treated as categorical due to what was available for the weighted cohort which used on UDS and eRehabData LOS categories. Duration of PTA was treated as a continuous variable calculated as the number of days between the date of injury and date of emergence from PTA, as determined by the first of 2 consecutive scores of 76 or greater on the Galveston Orientation Amnesia Test or 25 or greater on the Orientation Log (Jackson, Novack, & Dowler, 1998; Levin, O’Donnell, & Grossman, 1979). For cases discharged from inpatient rehabilitation still in PTA, this value was imputed using number of days from injury to inpatient rehabilitation discharge. Rehospitalizations were coded as “yes” if the patient was rehospitalized in the year following their discharge from inpatient rehabilitation and “no” if they were not rehospitalized during the first year postinjury. Cause of injury was determined from the external cause of injury code (E-Code) categories included in the International Classification of Diseases, Ninth Revision, Clinical Modification associated with the TBI, and included “Vehicular” (includes motor vehicle, motorcycle, bicycle, ATV, pedestrian vs. auto), “Violence-related” (includes gunshots, assaults with blunt instrument, other violence), “Falls,” and “Other” (including sports, hit by falling/flying object, and other unclassified incidents; ICD-9-CM, 2004).

Functional status during inpatient rehabilitation

Two functional measures selected from the TBIMS NDB included the FIM™ (FIM) and the Disability Rating Scale (DRS). The FIM comprises 18 items to measure functional independence (Granger, Hamilton, Keith, Zielezny, & Sherwin, 1986) with item scores ranging from 1 (total assistance) to 7 (complete independence); higher scores indicate greater functional independence. Two subscales are derived from the FIM, including the 13 item Motor subscale (scores range 13 to 91) and the five item Cognitive Scale (scores range 5 to 35). FIM scores at rehabilitation admission and at discharge were used for analysis. The DRS is an eight-item scale that assesses cognitive ability to manage activities of daily living, need for assistance or supervision, and potential employability (Rappaport, Hall, Hopkins, Belleza, & Cope, 1982). Scores range from 0 (no disability) to 29 (extreme vegetative state), with higher score indicating greater disability. DRS scores at admission and at discharge were used for analysis. Preinjury residence and discharge disposition was also used as an indicator of general functional status, and both variables were categorized as “Private residence,” “Nursing home,” “Adult home” (adult foster care, independent living, transitional living facility, assisted living, supported living, group home), “Hospital” (acute care, rehabilitation, and other), “Subacute care,” and “Other” (shelter for homeless, homeless).

Alcohol use

Alcohol use status was derived from the number of drinks per week self-reported from the individual at 1-year postinjury using the Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System (n.d.) and classified as “Abstaining,” “Light” (<3 drinks per week), “Moderate” (3-14 drinks per week or 3-7 for women), and “Heavy/Binge” (heavy is >14 drinks per week for men and >7 for women; binge drinking is four or more drinks for women and five or more drinks for men within 2 hr) based on guidelines for healthy use by gender (Corrigan, Bogner, Lamb-Hart, & Sivak-Sears, 2003; U.S. Department of Health and Human Services, 2018).

Outcomes Measures

Functional outcome

Global functional outcome was assessed with the 8-point GOS-E that measures the ability to live, work, travel, and socialize independently after TBI. The scale includes scores ranging from 1 (dead) through 8 (upper-good recovery), with a higher score indicating better global functional status. GOS-E is a well-validated measure, and has been shown to be strongly associated with outcome categories on numerous alternative functional scales (Levin et al., 2001; Nichol et al., 2011; Shukla, Devi, & Agrawal, 2011; Wilson, Pettigrew, & Teasdale, 2000). For the TBIMS NDB, the GOS-E is first measured starting at 1-year postinjury. Individuals who had expired by 1-year postinjury (identified with a GOS-E score of 1) were excluded from analysis as this study focused only on those surviving at least 1 year.

Driving status

The variable for driving status at 1-year postinjury was dichotomized as participants who drive independently or do not drive independently (rides with someone else, public transit, special bus, van service; no motorized transportation). Driving is considered a concrete measure of functional abilities when it is achieved. Approximately 50% of those with moderate-severe TBI return to driving within 2 years of injury based on TBIMS research (Novack et al., 2010). Those who return to driving exhibit greater life satisfaction (Novack et al., 2010) and community activity (Rapport, Bryer, & Hanks, 2008). The issue of driving is of particular importance to older individuals, for whom driving cessation is associated with decreased out of home activity (Marottoli et al., 2000), sharp decline in health (Edwards, Lunsman, Perkins, Rebok, & Roth, 2009), and depression (Chihuri et al., 2016).

Procedures

Patients were recruited during their inpatient hospitalization for a moderate or severe TBI. Once consent was obtained, data were collected from a combination of medical record documentation and questionnaires during the inpatient rehabilitation stay, with follow-up telephone interviews conducted at 1, 2, and 5 years postinjury and every subsequent 5-year interval.

Statistical Analysis

This study utilized a subset of the weighted data set derived from the TBIMS NDB that included cases age 50 years and older at the time of injury who were admitted for inpatient rehabilitation following a primary diagnosis of TBI between October 1, 2001, and December 31, 2010. The sample size of TBIMS NDB cases limited to these criteria was 2,336, representing a final U.S. population resulting from the weighting processes of 97,935 cases. This weighted data set was then used to calculate descriptive statistics of the sociodemographic, clinical, and functional characteristics at rehabilitation discharge. Means and standard deviations were calculated for continuous variables, while proportions were calculated for categorical variables. For all analyses, the cohort was stratified by age group using 10-year intervals starting at age 50 to allow for closer examination of comparative differences between the age groups. As the data were weighted to the population, the proportions for confidence intervals were extremely small, and therefore are not presented. Of the 2,336 cases in the TBIMS NDB meeting eligibility criteria, 168 had expired by 1-year postinjury and 317 cases were lost to follow-up, and thus, were excluded from analysis. The remaining 1,845 cases in the sample were reweighted using weights adjusted to the corresponding outcome of interest.

To evaluate predictors of GOS-E at 1-year postinjury across age groups, a separate data set was utilized that comprised valid cases followed at 1-year postinjury. Variables were grouped into categorical blocks based on the domains presented above (Sociodemographics, Injury Severity, Functional Status, and Alcohol Use) and entered into hierarchical regression models as progressive blocks in that same order. Block regression analysis was used to examine the added contribution of each variable block to the variance of GOS-E within each age group, while adjusting for covariates. The GOS-E was treated as a pseudo-continuous variable, to interpret any clinically meaningful change between levels. Adjusted R² values were used to report the effect sizes for each progressive model as variable blocks were added, as well as correct for any overestimation of the proportion of variance in GOS-E when generalized to the population. Statistical significance for these regression analyses were set at p < .01. These statistics were calculated using SPSS Version 21 (SPSS Inc., Chicago, Illinois).

A modified multivariate Poisson regression model with robust error variance was used to estimate the relative risk for driving ability at 1-year postinjury. This outcome was recoded as a dichotomous variable (yes/no). The same variable blocks were entered into progressive models, which were then compared for overall fit using the Akaike Information Criteria (AICs). Prevalence risk estimates of specific variables found to be statistically significant within each age group were calculated. Statistical significance for these regression analyses were also set at p < .01. These analyses were performed using STATA software program (Release 13).

Results

Sample

The sample of interest includes adults over age 50 who received inpatient rehabilitation for TBI. We categorized age at injury by decades (i.e., 50-59, 60-69, 70-79, and 80 or more years) to investigate differences across subgroups of older individuals rather than grouping all older adults together. The analyses presented below are based on the TBIMS weighted longitudinal NDB which represents a sample of 97,935 TBI cases admitted for inpatient rehabilitation between 2001 and 2010 in the United States.

Outcomes at Rehabilitation Discharge

Sociodemographic characteristics

Sociodemographic descriptive characteristics of the cohort are presented in Table 1. As these values are weighted to represent the U.S. TBI inpatient rehabilitation population, differences between age groups are assumed to be significant. Those aged 80 years and older accounted for the highest proportion of cases at 37%. The youngest group evaluated (those 50 to 59 years of age) was predominantly male (72.4%) and Caucasian (80.2%). The proportion of females increased with age, as the 80 years and older group had an almost equal gender distribution. Race and education maintained consistent proportions across the age groups, with the majority reporting being Caucasian and reporting a high school level education. The greatest shift of payor source as age increased was from private insurance to Medicare between the 50- to 59-year-olds and 60- to 69-year-olds groups, as would be expected with age 65 being the minimum age for general eligibility of Medicare benefits.

Table 1.

Incidence of Demographic, Clinical, and Functional Characteristics at Rehab Discharge by Age Group.

	Age groups
	50-59	60-69	70-79	80 and older	Total
	n = 16,896	n = 17,522	n = 27,535	n = 35,982	N = 97,935
	%	%	%	%	%
Gender*
Male	72.4	65.1	55.2	51.3	58.5
Female	27.6	34.9	44.8	48.7	41.5
Race*
Caucasian	80.2	80.8	79.5	87.2	82.7
Black	10.4	9.1	5.5	4.3	6.6
Hispanic	5.6	6.7	8.0	3.9	5.8
Other	3.7	3.4	7.0	4.6	4.9
Education* (% cases missing = 1.9)
11th grade or less	17.9	22.8	22.3	24.2	22.3
High school	38.5	34.1	36.6	36.1	36.3
Some college	21.9	17.0	15.9	12.8	16.0
College and postgraduate	21.7	26.1	25.2	26.9	25.4
Cause of injury* (% cases missing = 0.1)
Vehicular	50.2	30.5	26.8	11.4	25.8
Violence-related	9.7	4.5	1.5	0.9	3.2
Fall	37.3	61.0	70.8	87.1	69.3
Other	2.8	3.9	1.0	0.7	1.7
GCS group* (% cases missing = 6.0)
Severe	19.3	14.9	7.2	3.9	9.5
Moderate	16.0	11.5	8.6	4.2	8.9
Mild	44.6	58.3	79.8	88.4	72.8
Sedated	20.1	15.2	4.4	3.5	8.8
Length of stay (in days)*
1-9	27.5	30.0	28.6	25.8	27.6
10-19	41.3	40.0	41.6	44.7	42.4
20-29	18.2	19.6	20.1	20.1	19.7
30-39	6.3	5.0	5.6	5.8	5.7
40 and over	6.6	5.4	4.1	3.7	4.6
Primary payor*
Private insurance	50.9	31.2	16.8	8.3	22.1
Medicare	21.8	54.2	79.8	89.8	68.9
Medicaid	10.2	5.0	1.8	1.7	3.8
Workers comp	10.3	7.4	1.6	0.2	3.6
Self or no pay	6.8	2.2	0.0	0.0	1.6
Preinjury residence*
Private residence	96.4	98.7	94.9	94.4	95.7
Nursing home	0.0	0.0	0.2	1.0	0.4
Adult home	2.0	0.8	3.3	4.1	2.9
Hospital–any	0.2	0.0	0.6	0.2	0.3
Other	1.3	0.3	1.0	0.3	0.7
Discharge disposition*
Private residence	79.4	75.7	73.4	62.3	70.8
Nursing home	5.7	9.4	12.1	13.5	11.0
Adult home	2.8	2.4	3.1	5.7	3.9
Hospital–any	5.3	5.4	4.0	5.8	5.1
Subacute care	5.7	7.0	7.4	11.6	8.6
Other	1.0	0.0	0.0	1.1	0.6
N	16,785	17,391	27,474	35,757	97,407
	27.05 (28.53)	23.29 (27.20)	17.49 (21.11)	16.04 (19.98)	19.64 (23.72)

Note. GCS = Glasgow Coma Scale.

p < .001.

Clinical characteristics

Clinical and functional descriptive characteristics of the cohort are presented in Table 1. The predominant cause of injury was vehicular crashes among those aged 50 to 59, but the proportion of injuries resulting from falls increased with age, with the greatest proportion occurring within the 80+-year-old group (87%). TBI severity across all groups as measured by the GCS was predominantly categorized as “mild” (72.8%), followed by severe (9.5%), moderate (8.9%), and unknown due to chemical sedation (8.8%). The chemically sedated group was most similar to the severe GCS group in terms of other severity-related outcomes such as PTA and rehabilitation LOS. The largest proportion of “mild” TBI occurred among the 80 years and older group (88.4%), while the 50- to 59-year-olds group had the largest proportion of patients with severe (19.3%) or unknown due to sedation (20.1%) scores. The 80 years and older group also reported the lowest mean number of days in PTA at 16.0, compared to the 50- to 59-year-olds group with an average of 27.1 days in PTA, demonstrating a similar pattern of decreasing injury severity with increasing age. In addition, there were relatively similar proportions among the categories of total length of hospitalization across all age groups, with the largest proportion (ranging from 40% to 44%) of each age group reporting a total LOS of 10 to 19 days. Overall, TBI severity indicators presented a consistent pattern of decreasing severity with increasing age. When comparing preinjury residence and discharge disposition among the age groups, all age groups were predominantly living in a private residence prior to their TBI (94% to 98%). However, the proportion of individuals returning to a private residence after discharge decreased as age increased, implying greater dependence despite lesser TBI severity. The largest proportion being discharged somewhere other than private residence was observed among the 80 years and older age group (32%).

Functional status during inpatient rehabilitation

Mean scores on functional measures (FIM Motor, FIM Cognitive, and DRS) collected at admission and discharge across all age groups are shown in Figures 1 to 3. All comparisons between function and age groups reached statistical significance (p < .001) with all age categories improving during inpatient rehabilitation. Mean FIM motor subscores at both rehabilitation admission and discharge decreased with age, indicating less functional independence as age increased. Overall improvement in motor functioning between admission and discharge also decreased as age increased. Similar trends were noted for the mean FIM cognitive subscores and DRS. All age groups had relatively similar cognitive scores at the time of admission. However, by discharge, 50- to 59-year-olds had the largest improvement in cognitive functional independence (FIM cognitive) and level of disability (DRS) during inpatient rehabilitation, as compared with the older age groups. Improvement between admission and discharge was less as age increased.

Figure 1.

Mean FIM–motor at rehabilitation admission and discharge by age category.

Figure 2.

Mean FIM–Cognitive at rehabilitation admission and discharge by age category.

Figure 3.

Mean DRS score at rehabilitation admission and discharge by age category.

Outcomes at 1-Year Post-Injury

Extended Glasgow Outcome Scale (GOS-E)

Table 2 contains the model fit predicting GOS-E outcome at 1-year postinjury by age group. All multivariate regression models satisfied standard tests for normality, linearity, and multicollinearity. The full model including all predictors and covariates accounted for 24% of variance in GOS-E among the 50- to 59-year-olds, 34.3% of GOS-E variation among 60- to 69-year-olds and 70- to 79-year-olds groups and 25.4% variance among the 80 years and older group. The categorical block including demographic variables did not provide much explanatory power in the 50 to 59 and 60 to 69 groups, but did significantly contribute to the model for the 70 to 79 and 80+ groups. The clinical and functional blocks of covariates significantly improved the models across all age groups, with statistically significant increases in the adjusted R² values. The functional variable block contributed the most to the full model for all age groups, and this contribution was most evident among the 60- to 69-year-old group (adjusted R² = 0.32, ∆R² = 0.20) and 70- to 79-year-old group (adjusted R² = 0.33, ∆R² = 0.18). The addition of alcohol use in the full models did not contribute significantly to the overall GOS-E variance in any age groups, except the 70- to 79–year-olds (adjusted R² = 0.34, ∆R² = 0.02).

Table 2.

Model fit Predicting GOS-E Outcome at 1-Year Postinjury by Age Category.

Model:	1		2		3		4
Model:	Adjusted R²	ΔR²	Adjusted R²	ΔR²	Adjusted R²	ΔR²	Adjusted R²	ΔR²
50-59 (n = 269)	.034	.078	.115	.1*	.240	.139*	.240	.008
60-69 (n = 255)	.023	.069	.124	.12*	.320	.202*	.343	.028
70-79 (n = 385)	.059	.083*	.156	.11*	.327	.177*	.343	.02*
80 and older (n = 468)	.034	.053*	.151	.128*	.258	.116*	.254	.000

Note. Model 1 = demographic, Model 2 = demographic + clinical, Model 3 = demographic + clinical + functional status, Model 4 = demographic + clinical + functional status + drinking status. GOS-E = Glasgow Outcome Scale-Extended.

p < .001

Effect sizes and confidence intervals of significant variables (p ≤ 0.01) in each age group are provided in Table 3. Among adults age 50 to 59 years, rehospitalizations during the first year following injury was significantly associated with decreased global functioning. For adults age 60 to 69 years, clinical and functional variables were significant predictors of global function as well, with longer lengths of hospital stay and discharge disposition to a nursing home or assisted living facilities after inpatient rehabilitation having the largest negative effect on global functioning in this age group.

Table 3.

Multiple Regression Coefficients of Final Model By Age Category.

Age groups	Unstandardized coefficients		Standardized coefficients	95% CI
Age groups	B	SE	Beta*	Lower bound	Upper bound
80 and older
Hispanic	–1.366	.482	–.138	–2.313	–0.420
FIM motor at discharge	.033	.008	.223	0.017	0.048
FIM cognitive at discharge	.070	.021	.184	0.028	0.111
70-79
Hispanic	–1.070	.349	–.139	–1.755	–0.384
Days From acute admit to rehab discharge	–.018	.005	–.163	–0.029	–0.008
FIM motor at discharge	.027	.008	.198	0.010	0.043
Rehospitalized	–1.008	.202	–.229	–1.405	–0.612
Adult home	–1.575	.572	–.123	–2.699	–0.451
60-69
Days From acute admit to rehab discharge	–.011	.004	–.153	–0.020	–0.003
FIM motor at discharge	.023	.008	.213	0.006	0.039
Nursing home	–1.158	.379	–.169	–1.906	–0.411
Adult home	–2.080	.681	–.160	–3.422	–0.738
50-59
Rehospitalized	–.607	.223	–.155	–1.045	–0.168

Note. CI = confidence interval; FIM: Functional Independent Measure.

p < .001

Similar associations were found among the 70 to 79 group, with longer lengths of stay, any rehospitalizations and being discharged to an assisted living facility being significantly predictive of lower GOS-E outcome. Greater motor function at discharge was predictive of greater global functioning. For adults age 80 years and older, higher motor and cognitive functioning at discharge were also significant predictors of greater global functioning. For both these age groups, Hispanic ethnicity was associated with lower GOS-E scores.

Driving status

Table 4 shows the adjusted risk ratios (aRR) of explanatory variables for driving status at 1-year postinjury after accounting for all other covariates in the model. Model fit was evaluated using the AIC, with lower AIC indicating better fit. The results are summarized in Table 5. The addition of each progressive block reduced the AIC, with the final model reporting the lowest AIC value of 1.16. After controlling for all other variables, adults age 80 and older were 56% less likely to drive following TBI, which was significantly different than 50- to 59-year-olds. Adjusted risk ratios of age and driving were not found to be significantly different among the other age groups. Other significant predictor variables reported in the final model included sex, with females being 23% less likely to drive than males, after controlling for all other variables; ethnicity, with both Hispanic and African American individuals being 60% and 57% less likely to drive than white individuals, respectively. Rehospitalization during the first year following injury was also a significant predictor, with those who were rehospitalized 29% less likely to drive than those who were not rehospitalized. Those with drinking behavior described as either “light” or “moderate” were 53% and 44% more likely to drive than those who reported no drinking.

Table 4.

Adjusted Risk Ratio Estimates for Driving at 1-Year Postinjury.

	Adjusted risk ratio	SE	95% CI
	Adjusted risk ratio	SE	Lower bound	Upper bound
Age groups (ref = 50-59)
80 and older	0.44*	0.09	0.29	0.65
70-79	1.09	0.11	0.89	1.33
60-69	1.15	0.09	0.98	1.35
Gender (ref = Male)
Female	0.77*	0.07	0.64	0.93
Race (ref = White)
Other	0.75	0.15	0.50	1.11
Black	0.43*	0.08	0.30	0.62
Hispanic	0.40*	0.15	0.19	0.83
Education (ref = College/Postgrad)
11th grade or less	0.79	0.12	0.59	1.06
High school	1.01	0.10	0.82	1.23
Some college	1.02	0.11	0.82	1.27
Payor Source (ref = Private Insurance)
Other	1.00	0.19	0.68	1.46
Medicare	0.93	0.08	0.78	1.11
Medicaid	0.68	0.12	0.48	0.97
Workers comp	0.92	0.12	0.71	1.18
No Pay	1.17	0.32	0.68	2.01
Cause of injury (ref = Falls)
Vehicular	1.13	0.10	0.95	1.35
Violence-related	1.03	0.19	0.71	1.48
Other	1.66	0.43	1.00	2.75
GCS group (ref = Mild)
Sedated	1.06	0.10	0.88	1.28
Severe	0.96	0.11	0.77	1.20
Moderate	0.90	0.12	0.69	1.17
Rehospitalizations (ref = No)
Yes	0.71*	0.07	0.58	0.87
Discharge disposition (ref = Private Residence)
Other	0.72	0.60	0.14	3.68
Nursing home	0.76	0.13	0.54	1.07
Adult home	0.43	0.25	0.14	1.36
Hospital—Any	0.72	0.28	0.33	1.55
Subacute care	0.66	0.14	0.44	1.00
Drinking status (ref = Abstaining)
Heavy	1.15	0.26	0.74	1.79
Moderate	1.53*	0.16	1.24	1.89
Light	1.44*	0.15	1.18	1.75

Note. CI = confidence interval; GCS = Glasgow Coma Scale.

p < .01.

Table 5.

Estimate of Model Fit for Modified Poisson Regression Models for Likelihood of Driving at 1-Year Postinjury.

	Akaike information criteria
Model 1	1.175597
Model 2	1.179701
Model 3	1.170619
Model 4	1.164572

Discussion

The purpose of this study was to characterize population estimates of specific sociodemographic, clinical, functional, and lifestyle characteristics and to examine their associations with 1-year functional outcomes in older adults with TBI. By utilizing the TBIMS weighted longitudinal NDB to study these relationships in greater detail among a large and representative sample of older adults, this investigation provides further insight into patterns of long-term recovery among subgroups of older adults after TBI. This study extends the findings of prior studies through the inclusion of those 50 and older, dividing the older adults into smaller age categories, and the use of outcome measures that are meaningful to those living with TBI.

The reported estimates confirm the relevance of previous findings (Cuthbert et al., 2014; Dams-O’Connor et al., 2013; Marquez de la Plata et al., 2008) on a population level, demonstrating that older age is associated with milder injuries but less improvement in functional outcomes occurs, despite similar lengths of acute care and rehabilitation hospital stays across all age groups. The final models containing sociodemographic, clinical, functional, and alcohol use variable blocks made different contributions to global functioning at 1-year postinjury for each age group. After adjusting for sociodemographic and clinical variables, functional outcome at discharge had the greatest contribution in explaining variation in global functioning at 1 year across all older age groups, with the largest effect sizes observed among the 60- to 69-year and 70- to 79-years-olds age groups. This suggests that especially for older patients with TBI, maximizing functional improvement during inpatient rehabilitation may be essential to maximizing the potential for long-term recovery. It may be that older adults require augmented inpatient rehabilitation programs to achieve maximal gains during rehabilitation; this may involve greater care intensity and/or duration earlier in the acute recovery phase. This is an important consideration in a climate where inpatient rehabilitation LOSs are becoming briefer (Ratcliff et al., 2007; Barclay 2004). Care tailored more specifically to the needs of older adults with TBI may contribute to more strategic resource planning, thereby improving health care delivery, functional outcomes, and quality of life.

The different contributions of these variable blocks across age groups also suggest that some of the factors that are commonly associated with functional outcomes in younger adults are not as relevant for older adults. Alcohol use did not significantly contribute to global functioning for any group except those 70 to 79 years old. Similar to findings from recent studies (Corrigan et al., 2014) there were relatively lower rates of alcohol use among this TBI cohort than in the general TBI population. Despite relatively lower alcohol use in this study, excessive alcohol use is a targeted public health issue specific to older adults; in the general population, more than 14% of men and 3% of women age 65 years and older reported binge drinking behavior (Blazer & Wu, 2009). In addition, alcohol use can increase the risk of falls, have negative interactions with medications, and exacerbate or cause various chronic health problems.

The lack of association of driving with injury characteristics and functional measures during rehabilitation suggests that the injury itself was not a barrier to driving in this study population. After controlling for all other variables, female sex and any rehospitalizations during the first year were the greatest impediments to independent driving at 1 year following injury. Race, specifically being African American or Hispanic, was negatively associated with driving at 1 year postinjury. This is discouraging in that gender and race/ethnicity are immutable. However, there is no reason that TBI would have a greater impact on women or African Americans/Hispanics as related to driving. There are multiple factors that can influence return to driving, many of them socioeconomic, but to the extent that failure to drive reflects a divergence of opportunity, this needs to be addressed. Rehabilitation care providers can educate TBI patients in these groups and family members regarding the potential for return to driving when appropriate, while also promoting improved access to driving evaluations. There is evidence that following TBI, those who are not offered driving evaluation are less likely to drive (Rapport, Hanks, & Bryer, 2006). The association of driving with alcohol consumption is difficult to interpret other than as a measure of increased mobility and social contact afforded by driving.

In addition to alcohol use, closer examination of other potentially modifiable factors not included in this study such as, chronic comorbidities, and history of falls and/or prior TBIs may further our understanding of the extent that these factors affect long-term recovery among older adults. Studies have shown that frailty and gait instability can contribute to an increased risk of falls, which accounts for 51% of all reported TBIs among adults aged 65 and older (Thompson et al., 2006; U.S. Department of Labor, Bureau of Labor Statistics, 2014). A single fall is considered to be a major risk factor for both subsequent falls and subsequent TBIs (Thompson et al., 2006; Tinetti, 2003) making falls a relevant target for preventive health efforts.

Regarding Hispanic participants in this study and interpretation of their study findings, previous studies have reported negative impact of demographic, clinical, and functional characteristics of Hispanic individuals on post-TBI rehabilitation outcomes. These include acculturation, highest level of education, shorter length of rehabilitation stay, and FIM Cognitive Subscale score (potentially secondary to English language proficiency; Arango-Lasprilla et al., 2012; Ketchum et al., 2012; Lequerica et al., 2015). A more detailed examination of the impact of Hispanic ethnicity, preferably with a larger sample size in older age groups, is warranted to confirm results obtained in this study.

The study findings point to potentially unique rehabilitation needs and considerations following TBI among older adults. Meaningful gains were achieved during inpatient rehabilitation supporting the benefit of this level of care. We found less motor and cognitive improvement during rehabilitation as age increased, while total LOS did not differ across age groups. Prior studies have revealed older adults with TBI in general receive shorter rehabilitation LOS resulting in fewer hours of therapies (Dijkers, Brandstater, Horn, Ryser, & Barrett, 2013; Frankel et al., 2006), receive less psychology and therapeutic recreation services (Dijkers et al., 2013), less engagement in rehabilitation therapies (Seel et al., 2015), differences in medication administration (Hammond, Barrett, et al., 2015); greater risk of death (Brooks, Shavelle, Strauss, Hammond, & Harrison-Felix, 2015; Dijkers et al., 2013), more frequent rehospitalization that interrupts rehabilitation and rehospitalization subsequent to rehabilitation discharge (Dams-O’Connor et al., 2017; Hammond, Horn, Smout, Beaulieu, et al., 2015; Hammond, Horn, Smout, Seel, et al., 2015), and greater likelihood of institutionalization upon rehabilitation discharge and over time (Corrigan et al., 2014; Dijkers et al., 2013; Frankel et al., 2006). Medical and psychiatric comorbidities vary with age with diabetes, high blood cholesterol, osteoarthritis, and hypertension more common among older individuals with TBI (Hammond et al., 2019). Compared to younger individuals, deaths due to circulatory, fall, all respiratory, pneumonia causes, aspiration pneumonia, sepsis, nervous system (e.g., Alzheimer disease), and mental disorder (eg, dementia) related deaths have been found to occur more commonly for older individuals after TBI (Harrison-Felix et al., 2012). In the current study, we found that institutionalization and rehospitalization are related to worse outcome, suggesting that improvements in these areas would have meaningful impact on decreasing costs and improving outcomes. We should consider enhancing rehabilitation care for older adults with TBI to better meet their unique needs. For example, we should consider adjusting our expectations to allow longer LOS and less efficiency, customizing goals, individualizing treatment, involving family early, and encorporating means to enhance engagement with the goal of greater functional gains and less discharge to institutional settings. Psychology and therapeutic recreation services may have a role in enhancing engagement and improving outcome. With falls as the most common cause of TBI among older individuals, attention to fall prevention during rehabilitation is critical. This includes obtaining a thorough history, assessing and training balance, reviewing home setting, and improving motor function. Inpatient and outpatient care should also address polypharmacy, treat medical comorbidities, and aim to prevent medical complications common after TBI in older adults. Given the well-established relationship between age and rehospitalization, we should consider enhanced discharge planning and supported transitions in care, earlier postacute physician follow-up, risk-based case management and/or care plans. Mas, Mathews, and Gilbert-Bffoe (2017) offers suggestions for comprehensive rehabilitation for TBI rehabilitation in the elderly along with supporting rationale.

Study Limitations

There are some limitations with using weighted cohort data from the TBIMS NDB to derive population estimates. As previously noted, the characteristics of U.S. TBI inpatient rehabilitation population were slightly different than those meeting TBIMS NDB enrollment criteria when calculating the weights, which could have introduced bias in overestimating the effects of characteristics unique to the TBIMS NDB.

Limitations of generalizability include exclusion of those who do not receive inpatient rehabilitation, a largely Caucasian population, and the exploratory nature of this study. Since this is an investigation of people who received inpatient rehabilitation, it is unknown how generalizable these results may be for individuals with TBI who do not receive inpatient rehabilitation, such as those with mild TBI only treated in the emergency department, or those who were deemed to be too impaired to benefit from inpatient rehabilitation. The reported estimates excluded data from expired cases, which could have potentially introduced survivor bias in underestimating the incidence of various risk factors.

Analyses were also limited by the unavailability of certain variables that may have strengthened the explanatory power for predicting global functioning and driving status following TBI, such as preexisting chronic conditions, measures of current comorbidities, tobacco use, and current medications. Some of these variables have not been systematically collected for the TBIMS NDB; other variables, such as preexisting conditions and limitations, were added to the data collection in later years. LOS, analyzed here as the total days in acute care and rehabilitation, was not divided into acute care LOS and rehabilitation LOS which might yield different relationships with age. As the data set comprised cases collected from 2001 to 2010, inclusion of more recently collected variables in the analysis would have reduced the sample size by 50% or more, due to systematically missing data. However, future analyses may be able to include such variables. Future analyses could also benefit from expanding the time period for utilizing weights to include longer-term outcomes at 5 and 10 years postinjury.

Conclusion

The current study provided population estimates of sociodemographic, clinical, functional and lifestyle characteristics for adults aged 50 years and older who were admitted for TBI rehabilitation in the United States. Estimates of these factors, along with their relationships with outcomes reported here, provided broad scale validation of rehabilitation and recovery trends examined in previous studies. The finding that functional outcome at discharge from inpatient rehabilitation is the strongest contributor to better global functioning across all age groups has important implications for the care needs of older adults, which may include longer and more intensive acute rehabilitation to achieve maximal outcomes. Additional research into age-specific predictors of outcomes is warranted to inform the development of targeted interventions that can optimize TBI recovery in our aging population.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The contents of this publication were developed under grants from the National Institute on Disability, Independent Living, and Rehabilitation Research: Indiana University School of Medicine/Rehabilitation Hospital of Indiana (Grant #90DP0036), New York Traumatic Brain Injury Model System at the Icahn School of Medicine at Mount Sinai (Grant #90DPTB0009), Rocky Mountain Regional Brain Injury System (Grant #90DP0034), University of Alabama at Birmingham (Grant #90DP0044), Traumatic Brain Injury Model Systems National Data and Statistical Center (Grant #90DP0084), Spaulding/Harvard Traumatic Brain Injury Model System North Texas Traumatic Brain Injury Model System (Grant # 90DPTB0013). NIDILRR is a Center within the Administration for Community Living (ACL), Department of Health and Human Services (HHS). The contents of this publication do not necessarily represent the policy of NIDILRR, ACL, HHS, and you should not assume endorsement by the Federal Government.

ORCID iD

FM Hammond

References

Arango-Lasprilla

J. C.

Ketchum

J. M.

Drew

Hammond

Powell

J. M.

Kreutzer

. . . Carr

(2012). Neurobehavioural symptoms 1 year after traumatic brain injury: A preliminary study of the relationship between race/ethnicity and symptoms. Brain Injury, 26, 814-824.

Barclay

(2004). Inpatient rehabilitation length of stay is getting shorter. Journal of the American Medical Association, 292, 1687-1695.

Blazer

D. G.

L. T.

(2009). The Epidemiology of at-risk and binge drinking among middle-aged and elderly community adults: National survey on drug use and health. American Journal of Psychiatry, 166, 1162-1169.

Bogner

J. A.

Whiteneck

G. G.

MacDonald

Juengst

S. B.

Brown

A. W.

Phillippus

A. M.

. . . Corrigan

J. D.

(2017). Test-retest reliability of traumatic brain injury outcome measures: A Traumatic Brain Injury Model Systems Study. The Journal of Head Trauma Rehabilitation, 32(5), E1-E16.

Brooks

J. C.

Shavelle

R. M.

Strauss

D. J.

Hammond

F. M.

Harrison-Felix

C. L.

(2015). Long-term survival after traumatic brain injury part I: External validity of prognostic models. Archives of Physical Medicine and Rehabilitation, 96, 994-999.

Brown

A. W.

Malec

J. F.

McClelland

R. L.

Diehl

N. N.

Englander

Cifu

D. X.

(2005). Clinical elements that predict outcome after traumatic brain injury: A prospective multicenter recursive partitioning (decision-tree) analysis. Journal of Neurotrauma, 22, 1040-1051.

Centers for Disease Control and Prevention Behavioral Risk Factor Surveillance System. (n.d.). Atlanta, GA. Retrieved from https://www.cdc.gov/brfss/data_documentation/index.htm

Chihuri

Mielenz

T. J.

DiMaggio

C. J.

Betz

M. E.

DiGuiseppi

Jones

V. C.

(2016). Driving cessation and health outcome in older adults. Journal of the American Geriatrics Society, 64, 332-341.

Coronado

V. G.

Thomas

K. E.

Sattin

R. W.

Johnson

R. L.

(2005). The CDC traumatic brain injury surveillance system: Characteristics of persons aged 65 years and older hospitalized with a TBI. The Journal of Head Trauma Rehabilitation, 20, 215-228.

10.

Corrigan

Bogner

Lamb-Hart

Sivak-Sears

(2003). Technical report on problematic substance use variables. The Center for Outcome Measurement in Brain Injury. Retrieved from http://tbims.org/combi/subst/index.html

11.

Corrigan

J. D.

Cuthbert

J. P.

Harrison-Felix

Whiteneck

G. G.

Bell

J. M.

Miller

A. C.

. . . Pretz

C. R.

(2014). US population estimates of health and social outcomes 5 years after rehabilitation for traumatic brain injury. The Journal of Head Trauma Rehabilitation, 29(6), E1-E9.

12.

Corrigan

J. D.

Cuthbert

J. P.

Whiteneck

G. G.

Dijkers

M. P.

Coronado

Heinemann

A. W.

. . . Graham

J. E.

(2012). Representativeness of the traumatic brain injury model systems national database. The Journal of Head Trauma Rehabilitation, 27, 391-403.

13.

Cuthbert

J. P.

Corrigan

J. D.

Whiteneck

G. G.

Harrison-Felix

Graham

J. E.

Bell

J. M.

Coronado

V. G.

(2012). Extension of the representativeness of the Traumatic Brain Injury Model Systems National Database: 2001 to 2010. The Journal of Head Trauma Rehabilitation, 27(6), E15-27.

14.

Cuthbert

J. P.

Harrison-Felix

Corrigan

J. D.

Kreider

Bell

J. M.

Coronado

V. G.

Whiteneck

G. G.

(2014). Epidemiology of adults receiving acute inpatient rehabilitation for a primary diagnosis of traumatic brain injury in the United States. The Journal Head Trauma Rehabilitation, 30, 122-135.

15.

Dams-O’Connor

Cuthbert

J. P.

Whyte

Corrigan

J. D.

Faul

Harrison-Felix

(2013). Traumatic brain injury among older adults at level I and II trauma centers. Journal of Neurotrauma, 30, 2001-2013.

16.

Dams-O’Connor

Mellick

Dreer

L. E.

Hammond

F. M.

Hoffman

Landau

. . . Pretz

(2017). Rehospitalization over 10 years among survivors of TBI: A National Institute on Disability, Independent Living and Rehabilitation Research (NIDILRR) Traumatic Brain Injury Model Systems Study. The Journal of Head Trauma Rehabilitation, 32, 147-157.

17.

Dempster

A. P.

Laird

N. M.

Rubin

D. B.

(1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society. Series B (Methodological), 39, 1-38.

18.

Dijkers

Brandstater

Horn

Ryser

Barrett

(2013). Inpatient rehabilitation for traumatic brain injury: The influence of age on treatments and outcomes. Neurorehabilitation, 32, 233-252.

19.

Edwards

J. D.

Lunsman

Perkins

Rebok

G. W.

Roth

D. L.

(2009). Driving cessation and health trajectories in older adults. Journal of Gerontology: Medical Sciences, 64, 1290-1295.

20.

eRehabData. (2012). American Medical Rehabilitation Providers Association. Retrieved from https://web2.erehabdata.com/erehabdata/index.jsp

21.

Frankel

J. E.

Marwitz

J. H.

Cifu

Kreutzer

J. S.

Englander

Rosenthal

(2006). A follow-up study of older adults with traumatic brain injury: Taking into account decreasing length of stay. Archives Physical Medicine and Rehabilitation, 87, 57-62.

22.

Gardner

R. C.

Dams-O’Connor

Morrissey

M. R.

Manley

(2018). Geriatric Traumatic brain injury: Epidemiology, outcomes, knowledge gaps, and future directions. Journal of Neurotrauma, 35, 889-906.

23.

Goldstein

F. C.

(1998). Aging and recovery. Recovery, 9(3), 28-29.

24.

Goldstein

F. C.

Levin

H. S.

(1995). Neurobehavioral outcome of traumatic brain injury in older adults: Initial findings. The Journal of Head Trauma Rehabilitation, 10, 57-73.

25.

Goldstein

F. C.

Levin

H. S.

Goldman

W. P.

Kalechstein

A. D.

Clark

A. N.

Kenchan-Altonen

(1999). Cognitive and behavioral sequelae of closed head injury in older adults according to their significant others. Journal of Neuropsychiatry Clinical Neuroscience, 11, 38-44.

26.

Goldstein

F. C.

Levin

H. S.

Roberts

V. J.

Goldman

W. P.

Kalechstein

A. S.

Winslow

Goldstein

S. J.

(1996). Neuropsychological effects of closed head injury in older adults: A comparison with Alzheimer’s Disease. Neuropsychology, 10, 147-154.

27.

Goleburn

C. R.

Golden

C. J.

(2001). Traumatic brain injury outcome in older adults: A critical review of the literature. Journal of Clinical Geropsychology, 7, 161-187.

28.

Graham

J. E.

Radice-Neumann

D. M.

Reistetter

T. A.

Hammond

F. M.

Dijkers

Granger

C. V.

(2010). Influence of sex and age on inpatient rehabilitation outcomes among older adults with traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 91, 43-50.

29.

Granger

Hamilton

Keith

Zielezny

Sherwin

(1986). Advances in functional assessment for medical rehabilitation. Topics in Geriatric Rehabilitation, 1(3), 59-74.

30.

Hammond

F. M.

Barrett

R. S.

Shea

Seel

R. T.

McAlister

T. W.

Kaelin

. . . Horn

S. D.

(2015). Psychotropic medication use during inpatient rehabilitation for traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 96(8, Suppl.), S256-S273.

31.

Hammond

F. M.

Corrigan

J. D.

Ketchum

J. M.

Malec

J. F.

Dams-O’Connor

Hart

Whiteneck

G. G.

(2019). Prevalence of medical and psychiatric comorbidities following traumatic brain injury. Journal of Head Trauma Rehabilitation, 34, E1-E10.

32.

Hammond

F. M.

Horn

S. D.

Smout

R. J.

Beaulieu

C. L.

Barrett

R. S.

Ryser

D. S.

Sommerfeld

(2015). Readmission to an acute care hospital during inpatient rehabilitation for traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 96(8, Suppl.), S293-S303.

33.

Hammond

F. M.

Horn

S. D.

Smout

R. J.

Seel

R. T.

Beaulieu

C. L.

Corrigan

J. D.

. . . Brandstater

(2015). Rehospitalization during 9 months after inpatient rehabilitation for traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 96(8, Suppl.), S330-S339.

34.

Harrison-Felix

Kolakowsky-Hayner

S. A.

Hammond

F. M.

Wang

Englander

Dams-O’Connor

. . . Diaz-Arrastia

(2012). Mortality after surviving traumatic brain injury: Risks based on age groups. The Journal of Head Trauma Rehabilitation, 27(6), E1-E12.

35.

ICD-9-CM: The International Classification of Diseases, 9th Revision, Clinical Modification. (2004). Hyattsville, MD: National Center for Health Statistics.

36.

Jackson

W. T.

Novack

T. A.

Dowler

R. N.

(1998). Effective serial measurement of cognitive orientation in rehabilitation: The Orientation Log. Archives of Physical Medicine and Rehabilitation, 79, 718-720.

37.

Ketchum

J. M.

Getachew

M. A.

Krch

Baños

J. H.

Kolakowsky-Hayner

S. A.

Lequerica

. . . Arango-Lasprilla

J. C.

(2012). Early predictors of employment outcomes 1 year post traumatic brain injury in a population of Hispanic individuals. Neurorehabilitation, 30, 13-22.

38.

Kumar

R. G.

Olsen

Juengst

S. B.

Dams-O’Connor

O’Neil-Pirozzi

T. M.

Hammond

F. M.

. . . Wagner

A. K.

(2019). Comorbid conditions among adults 50 years and older with traumatic brain injury: Examining associations with demographics, health care utilization, institutionalization and 1-year outcomes. The Journal of Head Trauma Rehabilitation, 34, 224-232.

39.

Lequerica

Krch

Lengenfelder

Chiaravalloti

Arango-Lasprilla

J. C.

Hammond

F. M.

. . . Sander

A. M.

(2015). Comparison of FIMTM communication ratings for English versus non-English speakers in the traumatic brain injury model systems (TBIMS) national database. Brain Injury, 29, 1445-1451.

40.

Levin

H. S.

Boake

Song

McCauley

Contant

Diaz-Marchan

. . . Kotrla

K. J.

(2001). Validity and sensitivity to change of the extended Glasgow Outcome Scale in mild to moderate traumatic brain injury. Journal of Neurotrauma, 18, 575-584.

41.

Levin

H. S.

O’Donnell

V. M.

Grossman

R. G.

(1979). The Galveston orientation and Amnesia Test: A practical scale to assess cognition after head injury. Journal of Nervous and Mental Disease, 167, 675-684.

42.

Marottoli

R. A.

de Leon

C. F. M.

Glass

T. A.

Williams

C. S.

Cooney

L. M.

Jr. Berkman

L. F.

(2000). Consequences of driving cessation: Decreased out-of-home activity levels. Journal of Gerontology: Social Sciences, 55, S334-S340.

43.

Marquez de la Plata

C. D.

Hart

Hammond

F. M.

Frol

A. B.

Hudak

Harper

C. R.

. . . Diaz-Arrastia

(2008). Impact of age on long-term recovery from traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 89, 896-903.

44.

Mas

M. F.

Mathews

Gilbert-Bffoe

(2017). Rehabilitation needs of the elder with traumatic brain injury. Physical Medicine Rehabilitation Clinics of North America, 28, 829-842.

45.

Mazzucchi

Cattelani

Missale

Gugliotta

Brianti

Parma

(1992). Head-injured subjects aged over 50 years: Correlation between variables of trauma and neuropsychological follow-up. Journal of Neurology, 239, 256-260.

46.

Nichol

A. D.

Higgins

A. M.

Gabbe

B. J.

Murray

L. J.

Cooper

D. J.

Cameron

P. A.

(2011). Measuring functional and quality of life outcomes following major head injury: Common scales and checklists. Injury, 42, 281-287.

47.

Novack

T. A.

Labbe

Grotte

Carlson

Sherer

Arango-Lasprilla

J. C.

. . . Seel

R. T.

(2010). Return to driving within 5 years of moderate-severe traumatic brain injury. Brain Injury, 24, 464-471.

48.

2017 Profile of Older Americans. April 2018. Retrieved from https://acl.gov/sites/default/files/Aging%20and%20Disability%20in%20America/2017OlderAmericansProfile.pdf (accessed 6 December 2018).

49.

Ponsford

Cameron

Fitzgerald

Grant

Mikocka-Walus

(2011). Long-term outcomes after uncomplicated mild traumatic brain injury: A comparison with trauma controls. Journal of Neurotrauma, 28, 937-946.

50.

Rappaport

Hall

K. M.

Hopkins

Belleza

Cope

D. N.

(1982). Disability rating scale for severe head trauma: Coma to community. Archives of Physical Medicine and Rehabilitation, 63, 118-123.

51.

Rapport

L. J.

Bryer

R. C.

Hanks

R. A.

(2008). Driving and community integration after traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 89, 922-930.

52.

Rapport

L. J.

Hanks

R. A.

Bryer

R. C.

(2006). Barriers to driving and community integration after traumatic brain injury. The Journal of Head Trauma Rehabilitation, 21, 34-44.

53.

Ratcliff

J. J.

Greenspan

A. I.

Goldstein

F. C.

Stringer

A. Y.

Bushnik

Hammond

F. M.

Wright

D. W.

(2007). Gender and traumatic brain injury: Do the sexes fare differently? Brain Injury, 21, 1023-1030.

54.

Seel

R. T.

Corrigan

J. D.

Dijkers

M. P.

Barrett

R. S.

Bogner

Smout

R. J.

. . . Horn

S. D.

(2015). Patient effort in traumatic brain injury inpatient rehabilitation: Course and associations with age, brain injury severity, and time postinjury. Archives of Physical Medicine and Rehabilitation, 96(8, Suppl.), S235-S244.

55.

Shukla

Devi

B. I.

Agrawal

(2011). Outcome measures for traumatic brain injury. Clinical Neurology and Neurosurgery, 113, 435-441.

56.

Susman

DiRusso

S. M.

Sullivan

Risucci

Nealon

Cuff

. . . Benzil

(2002). Traumatic brain injury in the elderly: Increased mortality and worse functional outcome at discharge despite lower injury severity. Journal of Trauma, 53, 219-224.

57.

Thompson

H. J.

McCormick

W. C.

Kagan

S. H.

(2006). Traumatic brain injury in older adults: Epidemiology, outcomes, and future implications. Journal of the American Geriatrics Society, 54, 1590-1595.

58.

Tinetti

M. E.

(2003). Clinical practice. Preventing falls in elderly persons. The New England Journal of Medicine, 348(1), 42-49.

59.

Traumatic Brain Injury Model Systems National Database. (2015). Traumatic Brain Injury Model Systems National Data and Statistical Center. Available from http://www.tbindsc.org

60.

Uniform Data System for Medical Rehabilitation. (2012). Available from https://www.udsmr.org

61.

U.S. Department of Health and Human Services. (2018). Dietary guidelines for Americans. Retrieved from http://www.health.gov/dietaryguidelines/

62.

U.S. Department of Labor, Bureau of Labor Statistics. (2014). Unemployment: Rates and levels. Retrieved from http://www.bls.gov/data/#unemployment

63.

Utomo

W. K.

Gabbe

B. J.

Simpson

P. M.

Cameron

P. A.

(2009). Predictors of in-hospital mortality and 6-month functional outcomes in older adults after moderate to severe traumatic brain injury. Injury, 40, 973-977.

64.

Whiteneck

G. G.

Eagye

C. B.

Cuthbert

J. P.

Corrigan

J. D.

Bell

J. M.

Haarbauer-Krupa

J. K.

. . . Harrison-Felix

(n.d.). One and five year outcomes after moderate-to-severe traumatic brain injury requiring inpatient rehabilitation: Traumatic brain injury report. Retrieved from www.cdc.gov/TraumaticBraininjury/pdf/CDC-NIDILRR-Self-Report-508.pdf

65.

Wilson

J. T.

Pettigrew

L. E.

Teasdale

G. M.

(2000). Emotional and cognitive consequences of head injury in relation to the Glasgow outcome scale. Journal of Neurology, Neurosurgery, and Psychiatry, 69, 204-209.